PEMANFAATAN DATA PENJUALAN PRODUK SUSU BAYI PADA E-MARKETPLACE TOKOPEDIA DALAM PENENTUAN HARGA PRODUK DENGAN MENGGUNAKAN FRAMEWORK DYNAMIC PRICING

• Main Author: Adila, Dita
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/50626
• Institution: Institut Teknologi Bandung

Adskom is a company that provides services in the form of marketplace insights to infant and toddler formula milk companies selling their products on e-marketplaces such as Tokopedia. Currently, Adskom can obtain sales data from an average of 6.000 infant and toddler formula milk products on Tokopedia every day within five minutes. This data can be used by Adskom to assist their clients in the pricing process, which is currently being carried out by their clients on a trial-and-error basis and only based on competitor price benchmarks.. Better price management has the potential to increase company profits and revenues. One of the most suitable methods that can be applied in an e-marketplace environment is dynamic pricing. This research adopts the dynamic pricing framework developed by Bauer and Jannach (2018). In general, this framework is based on Bayesian inference combined with bootstrap-based confidence estimation and kernel regression. Specific historical sales data used as inputs to this framework are product name, product price, and the number of product sales. This framework yields an output in the form of the best price by considering competitors’ product prices. The best price for each product is the price that is predicted to achieve a certain profit and revenue targets. The calculation of root mean squared error (RMSE) and ????2 values in kernel regression and regression tree shows that these models have low prediction accuracy and these models cannot explain the variance in the model outputs quite well. This low level of framework performance is due to the limited data points used. Hence, it cannot be guaranteed that the model learning process is sufficient.